INTRODUCTION In order for an optical time-domain reflectometer OTDR to qualify as a candidate for complete calibration using this standard, it must be equipped with the following minimum
Organization
The calibration laboratory should satisfy requirements of ISO/IEC 17025
There should be a documented measurement procedure for each type of calibration performed, giving step-by-step operating instructions and equipment to be used.
Traceability
The requirements of ISO/IEC 17025 should be met
All calibration standards must be calibrated following a documented program that ensures traceability to national standards laboratories or accredited calibration laboratories It is recommended to have multiple standards at each hierarchical level to verify performance through comparisons Additionally, any test equipment that significantly impacts calibration results should also be calibrated, with its traceability chain specified upon request The re-calibration periods must be clearly defined and documented.
Preparation
All tests should be conducted at an ambient room temperature of 23 °C ± 3 °C, unless specified otherwise Ensure that the test equipment is allowed a minimum of 2 hours to acclimate to its environment before testing Additionally, the OTDR must undergo a warm-up period as per the manufacturer's instructions.
Test conditions
The test conditions usually include the following OTDR external conditions: date, temperature, connector-adapter combination and use of a lead-in fibre
Calibrate the OTDR according to the manufacturer's specifications and procedures Whenever possible, select a variety of test conditions and parameters that replicate the actual field operating conditions This selection should aim to enhance the OTDR's accuracy and resolution, utilizing features such as view windows and zoom, as outlined in the manufacturer's guidelines.
The test conditions usually include the following OTDR parameters: averaging time, pulse width, sample spacing, centre wavelength Unless otherwise specified, set the OTDR group index to exactly 1,46
NOTE 1 The calibration results only apply to the set of test conditions used in the calibration process
NOTE 2 Because of the potential for hazardous radiation, be sure to establish and maintain conditions of laser safety Refer to IEC 60825-1 and IEC 60825-2.
Documentation
Calibration certificates must include essential data and their associated uncertainties, such as the location offset \$\Delta L_0\$ and its uncertainty \$\pm 2u_{\Delta L_0}\$, along with the distance scale deviation \$\Delta S_L\$ and its uncertainty \$\pm 2u_{\Delta S_L}\$, or the location deviations \$\Delta L_i\$ and their uncertainties \$\pm 2u_{\Delta L_i}\$ Additionally, the certificate should specify the non-linearity (NL) loss, the instrument configuration (including pulse width, measurement span, wavelength, and averaging time) used during calibration, and any other relevant calibration data and requirements as stipulated.
General
The objective of distance calibration is to determine deviations (errors) between the measured and actual distances between points on a fibre, and to characterize the uncertainties of these deviations
An Optical Time Domain Reflectometer (OTDR) determines the location \( L \) of a feature by measuring the round-trip transit time \( T \) of a light pulse traveling to the feature and back The distance \( L \) is calculated using the speed of light in a vacuum, \( c \) (approximately \( 2.998 \times 10^8 \) m/s), and the group index \( N \) of the fiber.
Errors in measuring length (L) can arise from scale inaccuracies, timebase offsets in the Optical Time Domain Reflectometer (OTDR), and difficulties in accurately locating features relative to the timebase The placement of markers for measurement can be performed either manually or automatically by the instrument Generally, the level of error is influenced by the method of marker placement and the nature of the feature being measured, such as point losses, large reflections that saturate the receiver, or small reflections that do not.
Even larger errors in measuring L may result from the uncertainty in determining the multimode fibre's group index N and taking into account the differential mode delay The determination of
The analysis of differential mode delay and its consequences is not covered by this standard Therefore, the calibration procedures focus solely on the OTDR's capability to accurately measure T For this standard, a default value of N = 1.46 is adopted, with an uncertainty of 0 Additionally, the calibration methods are designed to minimize uncertainties related to differential mode delay.
Location deviation model
To characterize location deviations in Optical Time Domain Reflectometers (OTDRs), we will assume a specific model that reflects their typical behavior The reference location of a feature, denoted as \$L_{ref}\$, is measured from the front panel connector of the OTDR, while the displayed location is represented as \$L_{otdr}\$ It is assumed that the displayed location \$L_{otdr}\$, which utilizes OTDR averaging to reduce noise, functionally depends on the reference location \$L_{ref}\$.
The scale factor \( S_L \) should ideally be 1, while the location offset \( \Delta L_0 \) should be 0, and the distance sampling error \( f(L_{\text{ref}}) \) should also be 0 This distance sampling error is a periodic function with a mean of zero, characterized by a period that corresponds to the distance interval between sampled points on the OTDR For instance, when measuring the location of a significant reflection by marking the first digitized point that indicates an increase in signal, and then incrementing the position of the reflection in fine steps, the function \( f(L_{\text{ref}}) \) may resemble a periodic ramp waveform.
Equation (9) addresses the known errors in location measurements; however, an additional additive uncertainty type A may still exist This uncertainty impacts both the distance measurements and the precision of the parameters that describe these errors, as outlined in the following procedures.
To determine the values of \( S_L \) and \( \Delta L_0 \), one can measure \( L_{otdr} \) for various \( L_{ref} \) values and then apply the least squares method to fit a straight line to the data In this context, \( S_L \) represents the slope, while \( \Delta L_0 \) denotes the intercept of the fitted line.
Equivalently, a line may be fitted to the location deviation function, that is the difference between L otdr and L ref
L= =Δ ⋅ +Δ + Δ (10) where Δ S L is the slope, and Δ L 0 is still the intercept, as illustrated in Figure 2
The distance sampling error, denoted as f(L ref), can be determined by measuring deviations from the linear approximation for various values of L ref The amplitude of the distance sampling error, Δ L sample, is defined as half the amplitude of f(L ref).
In this standard, the amplitude of the distance sampling error, denoted as Δ L sample, is considered a component of the location readout uncertainty type A Consequently, the reported uncertainty result does not account for the repetitive nature of the sampling error, failing to differentiate between the contributions of the sampling error and uncertainty type A.
The distance calibration results are defined by several key parameters: the distance scale deviation (\$Δ S L\$) and its uncertainty (\$u Δ SL\$), the location offset (\$Δ L 0\$) and its uncertainty (\$u Δ L0\$), as well as the location readout uncertainty (\$u Lreadout\$) The latter represents the combined uncertainty arising from the distance sampling error and the type A uncertainty of the measurement samples, expressed as a standard deviation.
To determine the value of \$u_{Lreadout}\$, divide the maximum deviations from the least-squares approximation by the square root of 3, adhering to the mathematical basis It is important to note that the uncertainty is influenced by factors such as distance, the power level displayed, and the settings of the instrument.
NOTE Δ L sample represents the physical sampling error of the instrument This error is accessible for the user as u Lreadout that includes distance calculation and displaying errors.
Using the calibration results
The error in the location of a feature Δ L = L otdr – L ref can be calculated from the calibration results:
L=Δ + Δ Δ (11) with the uncertainty in ΔL given by the following formula, in which the recommended confidence level of 95 % is used:
2u =± u +L u +u ± (11a) where the displayed location L otdr can be used instead of the reference location L ref without serious consequences
Similarly, the error in the distance between two features ΔD and its uncertainty can be calculated from the following formula:
D= Δ Δ (12) with uncertainty in ΔD given by the following formula:
2u =± D u + u ± (12a) where the displayed distance D otdr can be used instead of the reference distance D ref
NOTE The 2 in front of u Lreadout 2 is due to combining two uncorrelated uncertainties
Differential mode delay may create additional uncertainties on long fibres measurement Such uncertainties should be negligible for distance given in Table 1
Length of fibre causing additional distance uncertainty
When calibrating, it's essential to consider additional uncertainties if the feature type differs from the one used in the calibration process Clearly specify the feature type in the calibration results to ensure accuracy and reliability.
Measuring fibre length
One effective method for OTDR distance calibration involves measuring fibers of known length This standard emphasizes determining fiber length through transit time rather than mechanical measurements, aligning with the OTDR's measurement principles Transit time measurements typically offer greater accuracy, especially for longer fibers Consequently, it is recommended to prioritize fiber transit time over mechanical length when precision is crucial.
To measure the transit time of the fiber, T transit, utilize a pulse generator, a triggerable laser source, an optical-to-electrical converter (O/E converter), and a time interval counter It is crucial that the laser source's center wavelength, λ center, closely matches that of the test OTDR, as any wavelength discrepancy can lead to variations in transit time due to chromatic dispersion Alternatively, the OTDR can be employed to generate optical pulses, ensuring that the center wavelengths align Record the transit time by calculating the difference in arrival times with and without the fiber positioned between the laser source and the O/E converter.
When this fibre is used for OTDR distance calibrations, then the reference distance D ref can be calculated by
In this equation, use a group index N which is identical with the OTDR's group index setting
The time measurement principle makes it possible to use D ref as the reference distance
General
Each of the calibration methods described below is capable of determining all of the necessary calibration results: location offset, distance scale deviation, and their uncertainties.
External source method
Short description and advantage
The external source method uses a calibrated time-delay generator to simulate the time delay in a fibre and an optical source to simulate the reflected or scattered signal from a fibre
To minimize uncertainties from differential mode delay, IEC 60793-2-50 single mode fibers are preferred over multimode fibers for interconnections, especially when operating at 1,300 nm.
This method is ideal for automated laboratory testing managed by computer control, focusing solely on reflective features for simplicity To calibrate the Optical Time Domain Reflectometer (OTDR) for non-reflective features, the pulsed electro-optical (E/O) converter should be substituted with an optical source that accurately simulates the desired feature.
Equipment
The measurement equipment, as illustrated in Figure 3, comprises several key components: a mode conditioner, a single mode optical coupler, an optical-to-electrical converter, a digital delay generator with pulse capability, an electrical to optical converter, and a variable optical attenuator to reduce the pulse amplitude just below the clipping level.
Figure 3 – Equipment for calibration of the distance scale –
The OTDR connects to the coupler via a mode conditioning multimode to single mode adapter, allowing the coupler to direct the OTDR signal to the O/E converter (detector) This detector activates the delay generator, which, after a predetermined time delay, generates an optical pulse that is subsequently coupled back to the OTDR.
The E/O converter functions as a pulsed laser that mimics reflection, utilizing constant pulse amplitude and width to effectively calibrate the distance scale for reflective features Additionally, the attenuator allows for the adjustment of pulse amplitude according to the distance of the reflection from the OTDR's front panel, simulating the variations in reflection amplitude due to fiber attenuation.
To allow accurate calibration of the set-up, fibres F1 and F5 should have the same length (see below) Fibre F5 is terminated to absorb reflections
NOTE 1 The mode conditioner is needed to make sure the OTDR receives proper launch conditions from the electrical to optical converter Therefore fibre F0 should be connected to the output of the mode conditioner while fibre F1 should be connected to the input
NOTE 2 The attenuation of the optical path between the connector of the OTDR and the optical to electrical converter may be high This is acceptable as the output power of the OTDR is generally sufficient
Calibration of the equipment
Before utilizing the "external source" equipment, proper calibration is essential It is important to ensure that the digital delay generator is regularly calibrated To compute the location offset \$\Delta L_0\$ from the measured data, the insertion delay \$T_{\text{delay}}\$ of the apparatus must also be determined This can be achieved by incorporating a pulse generator and a calibrated time interval counter into the setup, as illustrated in Figure 4.
F0 multimode fibre (two during calibration)
MC mode conditioner (two during calibration)
Figure 4 – Set-up for calibrating the system insertion delay
To properly measure the propagation delay of the mode conditioner it is recommended to include within the optical path, a second identical mode conditioner
To calibrate the insertion delay T delay , proceed as follows
To measure delay time accurately, configure the pulse generator to produce a square wave with a repetition period exceeding twice the delay time Utilize the output pulse from the generator as the start pulse for the time interval counter and to externally trigger the delay generator Ensure the digital delay generator is set for external triggering with zero delay aligned to the leading edge of the pulse generator signal, and adjust the trigger levels for both the delay generator and the counter accordingly.
To minimize uncertainty in the time interval measurement, it is crucial that the electrical cables E3 and E4, as well as the fibres F1 and F5, are of equal length Additionally, the two modes conditioner and fibres F0 must also match in length It is important to note that identical cable numbers in Figures 3 and 4 refer to the same physical cables For optimal triggering of the time interval counter, adjust the optical attenuator accordingly and record the displayed time interval, referred to as the insertion delay T delay.
Measurement procedure
Choose between automatic or manual techniques for feature location on the OTDR Set the attenuator to achieve the required pulse amplitude For instance, select a pulse width of 1 μs on the digital delay generator.
To optimize the delay generator T i settings, select time intervals that ensure samples are spread across a broad distance range with some randomness, facilitating averaging over the OTDR's distance sampling interval The initial time setting should position the pulse near the OTDR's front panel while remaining outside the initial dead zone for accurate measurements If the testing laboratory does not provide an alternative distance sampling scheme, one of the two specified methods must be utilized In the first method, assess the sample spacing D sample based on the appropriate OTDR instrument settings, such as by zooming into the OTDR trace, and subsequently calculate the corresponding delay difference of the delay generator T sample using the speed of light.
= (14) where N is the OTDR's group index setting and c is the speed of light in vacuum
Then calculate a total number of i delay generator settings, grouped in k clusters of n settings each (i = k n), where each cluster uniformly covers one sample spacing Each cluster shall have the form: n n T n T
In the first scheme, the clusters are defined as T K, K + sample, K + 2 sample, , K + (−1) sample, with each cluster containing at least four settings that are uniform across all clusters The cluster centers are evenly distributed, extending from just beyond the initial dead zone to a considerable distance for instrument calibration, with a minimum of two clusters In the second scheme, clusters are absent, and the sample spacing, D sample, only needs to be roughly estimated T sample can be calculated using Equation (14), with time settings uniformly spaced between the initial dead zone and a significant distance, each incorporating a random time interval These random intervals should follow a uniform probability density within the range of –T 1 to T 1, where T 1 is at least 20 times T sample but less than 10% of the longest time delay for the tests.
(that is, different settings) should be at least 20
Prior knowledge of the magnitude of type A uncertainty and acceptable measurement tolerances can guide testing laboratories in choosing an appropriate systematic or random distance sampling scheme.
Begin by selecting the initial time setting, T_i, for the series T_1 Document the delay generator's time, T_1, along with the measured location, L_otdr,1, of the event on the OTDR Follow the time settings outlined in section 6.2.4.1, ensuring to consistently record the time T_i and the corresponding measured location L_otdr,i Continue this process until all time settings have been completed.
Calculations and results
Following the concept of Clause 4, use the time settings to calculate i reference locations L ref,i
N is the group index setting of the OTDR;
T i are the time settings defined in 6.2.3;
T delay is the calibrated insertion delay of the test equipment (see 6.2.2)
Then, use the reference locations and the displayed locations L otdr,i to calculate the set of i location deviations Δ L i Δ L i = L otdr,i – L ref,i (17)
To calculate the location offset \$\Delta L_0\$ and the distance scale deviation \$\Delta S_L\$, the location deviation data is fitted to a simplified model that temporarily disregards distance sampling errors The model is expressed as \$\Delta L_i, \text{model} = \Delta S_L L_{\text{ref},i} + \Delta L_0\$.
Specifically, minimize the difference between the model and the data using the least-squares criterion that is, choose Δ S L and Δ L 0 so that the summation
∑ Δ L i − Δ S L − Δ L (19) is minimized Record Δ L 0 and Δ S L obtained from the approximation
The slope of the linear approximation, as illustrated in Figure 2, indicates the distance scale deviation, denoted as \$\Delta S_L\$ Meanwhile, the intercept on the vertical axis signifies the location offset, represented as \$\Delta L_0\$ It is essential to document the calculated values of \$\Delta S_L\$ and \$\Delta L_0\$.
Uncertainties
A general discussion of the distance uncertainties can be found in Clause 5
The list of uncertainties provided may not be exhaustive, and additional factors should be considered based on the specific measurement setup and procedure It is essential to utilize the mathematical framework outlined in Annex B for calculating and reporting these uncertainties.
The least-squares approximation described in section 6.2.5 utilizes the displayed distances between measurement samples to determine the distance scale deviation It is assumed that samples close to L = 0 and the maximum location L = L max significantly impact the distance scale deviation, while samples in the middle range exert less influence on the slope of the distance error model.
Applying the standard formula for the propagation of errors to Equation (4) yields the distance scale uncertainty u Δ SL in which ≅ D ref was used
D otdr is D ref ≈ L ref (for the long distances discussed here); u is the standard deviation expressing the uncertainty of the distance samples
The slope uncertainty, denoted as \( u / \), arises from inaccurate distance measurements and is equivalent to the standard deviation of the slope, \( \Delta S_L \), in the location model described by Equation (10) This uncertainty encompasses both marker placement inaccuracies and distance sampling errors The least-squares algorithm employed to calculate \( \Delta S_L \) can also be utilized to determine \( u \) If relevant, \( \Delta L_i \) may be averaged over the corresponding sampling interval Additionally, \( u Dref \) signifies the uncertainty of the reference distances, while \( u Dref /D ref \) indicates the slope uncertainty introduced by the digital delay generator, reflecting the relative timing uncertainty of the generator.
The location offset \$\Delta L_0\$ is determined by the intercept of the least-squares approximation on the vertical axis This intercept is primarily influenced by the initial samples near the location \$L = 0\$ and the precision of the insertion delay \$T_{\text{delay}}\$.
The location offset uncertainty u Δ L0 can be calculated by using the standard formula for the propagation of errors
The uncertainty of the differences between the measured values (\$Δ L_i\$) and the least-squares approximation near \$L = 0\$ is represented by \$u_Δ L\$, which accounts for marker placement uncertainty and distance sampling error This uncertainty is equivalent to the standard deviation of the differences between the measured and model values near \$L = 0\$ If applicable, the measured values can be averaged over the corresponding sampling interval The least-squares algorithm used to determine \$Δ L_0\$ can also be applied to calculate \$u_Δ L\$ Additionally, \$u_{Tdelay}\$ represents the uncertainty of the system insertion delay, which includes the differences between the two mode conditioning adapters used during calibration It is assumed that the first measurement setting will be very short or even zero, thereby reducing the delay generator uncertainty to one of the insertion delays only.
According to Clause 5, identify the maximum difference between the location deviation samples \$\Delta L_k\$ and the least-squares approximation at \$L = 0\$ To calculate the location readout uncertainty \$u_{L\text{readout}}\$ (which accounts for the distance sampling error), divide the maximum difference by the square root of 3 Alternatively, \$u_{L\text{readout}}\$ can also be calculated using the least-squares algorithm for determining \$\Delta S_L\$ and \$\Delta L_0\$ or by applying the specified formula.
Concatenated fibre method (using multimode fibres)
Short description and advantages
This method uses multimode calibrated fibres with transit times precisely measured at the wavelength of the OTDR under test to calibrate the distance scale
This method utilizes connectorized lengths of fiber, making it a cost-effective solution for testing in locations where traditional equipment is impractical Although it is primarily a manual testing approach that involves repeatedly connecting and disconnecting short fiber lengths to adjust reflection locations, automation is possible through the use of optical switches.
This technique utilizing multimode fibers is effective at both 850 nm and 1300 nm without the need for a mode conditioner However, it is essential to consider the differential mode delay (DMD) of the calibrated fiber B as a source of uncertainty.
Equipment
The equipment features a test OTDR along with several components: fibre A for identifying location offsets, fibre B for measuring distance scale deviations, and a set of incremental fibres to assess distance sampling errors, as illustrated in Figure 5.
One fibre of the set
Figure 5 – Concatenated fibres used for calibration of the distance scale
Normally, these fibres will be cabled or packaged in some way for protection and connectorized for easy connection and disconnection
Fibre A is a simple multimode fibre characterized by an end reflection from a non-angled connection Its shorter length compared to fibre B is not critical, provided it allows the end reflection to be measured on a backscatter trace that remains largely undisturbed by the initial reflections near the OTDR port.
Fibre A serves as a lead-in fibre for assessing distance scale deviation using Fibre B, which should feature reflective ends, such as those found in its connectors To minimize uncertainty, it is advisable for Fibre B to be several hundred meters in length.
To calibrate the fibre, measure its optical transit time T b as described in Clause 5
For accurate distance calibration, it is crucial that the reflections from both ends of fibre B (connectors C2 and C3) are approximately equal Discrepancies in reflections, such as one end saturating the OTDR while the other does not, can result in inaccurate distance measurements Although the impact of this difference on distance scale deviation is minimal for long fibres, it remains significant Additionally, when the manufacturer's sampling interval is uncertain and expected to be large, incremental fibres will be utilized to adjust the locations of the reflections from fibre B by amounts smaller than the distance sampling interval.
When using an OTDR, it is essential to choose fiber lengths that create at least four evenly spaced distance increments within the sampling interval For instance, with a distance sampling interval of 10 m, selecting fiber lengths of 2.5 m and 5 m can effectively achieve this goal These lengths can be utilized both separately and in combination to generate the required increments.
0 m (no fibre), 2,5 m, 5 m, and 7,5 m (both fibres) More generally, the fibres should generate length increments of
0 D D n− D (23) where n ≥ 4, and n D x equals the distance sampling interval of the OTDR under the conditions to be tested
In most cases, it is unnecessary to calibrate the transit time of these fibers as outlined in Clause 5; instead, measuring their physical lengths is sufficient The discrepancy between the true group index and the OTDR group index setting is minimal for such short fibers.
Measurement procedures
Random noise typically has a minimal impact on location deviation, unless the displayed power level approaches the instrument's noise limit In such situations, it is advisable to use longer averaging with the OTDR.
Establish the technique (automatic or manual) for placing the markers on the reflection of fibre
A and the reflective ends of fibre B
To visualize the end reflection of fibre A on the OTDR, connect fibre A to the front of the device Next, attach fibre B to the far end of fibre A, allowing the OTDR to capture reflections from both ends of fibre B.
Utilize the OTDR to measure the reflection point of fibre A, noting this initial measurement as \$L_{otdr,1}\$ Next, measure the length of fibre B with the OTDR, capturing the two reflections produced by this fibre, and record this distance as \$D_{otdr,1}\$.
Insert the shortest of the incremental fibres between the OTDR and the beginning of fibre A
Measure the location L otdr,2 and the distance D otdr,2
Continue inserting successively increasing length combinations of the incremental fibres
Measure the location L otdr,i and the distance D otdr,i until i = n and the total length of the incremental fibres is (n – 1) D x
Calculations and results
Compute the distance < D otdr > (the length of fibre B) as the average of the n values of D otdr,i
Then compute the distance scale deviation as
D ref is the reference distance;
N is the group index setting of the OTDR;
T b is the one-way transit time for fibre B, as measured according to Clause 5
Let < L otdr > be the average of all n values of L otdr,i Compute the location offset on the basis of
< L ref > is the average reference location corresponding to the first reflection, to be calculated with the help of the average length of the incremental fibres;
N is the group index setting of the OTDR;
The one-way transit time for fibre A, denoted as \$T_a\$, is measured in accordance with Clause 5 The distance scale deviation, represented by \$\Delta S_L\$, is calculated using Equation (10) If fibre A is sufficiently short, the term \$\Delta S_L\$ can be ignored.
Uncertainties
Clause 5 provides a comprehensive overview of distance uncertainties, highlighting that the listed uncertainties may not be exhaustive It is important to consider additional factors that could influence measurements based on the specific setup and procedures used.
The mathematical basis given in Annex B should be used to calculate and state the uncertainties
The distance scale uncertainty u Δ SL should be calculated with the following formula which is derived from Equation (10)
T D u u u m/km (27) where u Dotdr is the uncertainty of the displayed length of fibre B, for example, as caused by the marker placement uncertainty and the distance sampling error
The transit time uncertainty of fibre B, denoted as \( u_{Tb} \), is calculated using the root-sum-squaring method This includes several components: \( u_{Tb,counter} \), which accounts for the uncertainty from the time interval counter; \( u_{Tb, \lambda} \), representing the uncertainty due to the difference between the wavelength used for transit time measurement and the OTDR wavelength; \( u_{Tb, \Theta} \), which is influenced by the temperature coefficient of fibre B, typically valued at 1 cm/(km °C); and \( u_{Tb,DMD} \), reflecting the uncertainty from differential mode delay.
1 gives some indications for different multimode fibres
The location offset uncertainty u Δ L0 should be calculated from the following formula which is derived from Equation (25), by neglecting the Δ S L and the (n–1) D x /2 terms:
The uncertainty in measuring the location of the end reflection of fibre A, denoted as \$u_{Lotdr}\$, primarily arises from the placement of the marker It is assumed that any distance sampling error is effectively mitigated by averaging over a single sampling interval.
The transit time uncertainty of fibre A, denoted as \( u_{Ta} \), is calculated using the root-sum-squaring method This includes several components: \( u_{Ta,counter} \), which represents the uncertainty from the time interval counter; \( u_{Ta, \lambda} \), the uncertainty arising from the difference between the wavelength used for transit time measurement and the OTDR wavelength; \( u_{Ta, \Theta} \), which accounts for the temperature coefficient of fibre A, typically valued at 1 cm/(km °C); and \( u_{Ta,DMD} \), the uncertainty due to differential mode delay in fibre A.
(can be estimated based on the length ratio of the different type of fibres; should be negligible if fibre A is much shorter than fibre B)
Calculate the following two sets of data, one for the location deviation and one for the distance error, using the measurement samples given in 6.3.3
To calculate the location readout uncertainty \$u_{L\text{readout}}\$ for the L set or D set, it is advised to take half the difference between the largest and smallest values of the larger set and divide it by the square root of 3.
Recirculating delay line method
Short description and advantages
The recirculating delay line method uses a calibrated loop of multimode fibre, made with a coupler and a reflector, to generate periodic reflections
The technique resembles the concatenated fibre method, utilizing a fibre artifact without the need for electronic devices This artifact produces numerous calibrated distance samples, which can significantly minimize type A uncertainties that impact the deviation of the distance scale.
The measurements of location offset are limited to the reflective features generated by the recirculating delay line.
Equipment
In addition to the test OTDR, the measurement equipment only includes a multimode recirculating delay line manufactured and calibrated according to Annex A, as shown in Figure 6
Lead-in length (transit time T a)
Figure 6 – Distance calibration with a recirculating delay line
The recirculating delay line on the OTDR display features multiple reflective elements, as illustrated in Figure 7 The first element corresponds to the optical pulse traveling directly to the mirror and returning to the OTDR The second element is created by the optical pulse that travels once through the loop before reaching the mirror and returning to the OTDR, coinciding with the pulse that travels directly to the mirror The third pulse involves the optical signal traveling through the loop twice, and this pattern continues with additional reflections.
Accordingly, the ideal displayed locations would be etc
(31) where L a is the length of the lead-in fibre and L b is the length of the fibre loop
Di sp la yed power F ( dB)
Figure 7 – OTDR trace produced by recirculating delay line
Incorporating one or more incremental fibres into the measurement setup can be beneficial, although the necessity is diminished due to the averaging effect caused by multiple reflections from the delay line over the distance sampling interval This uncontrolled averaging may lead to a preference for systematic control According to the notation provided, setting \( n = 2 \) may suffice, allowing for a single incremental fibre with a length equal to half of the distance sampling interval.
Measurement procedure
The procedure operates under the assumption that no incremental fibers are utilized However, if incremental fibers are incorporated, the number of recorded distance samples increases, allowing for easy adjustments to the notation and calculations Consequently, the method closely resembles that of section 6.3, where the lead-in fiber corresponds to fiber A and the loop length is analogous to fiber B.
Establish the technique (automatic or manual) of placing the markers at the leading edges of the reflections from the recirculating delay line, following the manufacturer's recommendations
Connect the recirculating delay line assembly directly to the OTDR so that the reflective features can be seen on the OTDR
Measure the positions of consecutive reflections from the recirculating delay line using the OTDR, recording them as \( L_{\text{otdr},i} \), where the index \( i \) ranges from 0 to \( k \), indicating the number of passes through the loop While a higher value of \( k \) is expected to enhance result accuracy, it is ultimately constrained by loss and the OTDR's noise floor.
Calculations and results
Using the calibration data of the recirculating delay line T a and T b the series of reference locations is i = 1:
= etc where N is the group index setting of the OTDR
Then use the displayed locations L otdr,i and the reference locations to calculate the series of location deviations Δ L i:
To calculate the location offset \$\Delta L_0\$ and the distance scale deviation \$\Delta S_L\$, it is essential to fit the location deviation data to a simplified model that temporarily disregards the uncertainty in location readouts.
Specifically, minimize the difference between the model and the data using the least-squares criterion, that is choose Δ S L and Δ L 0 so that the summation:
∑ Δ L i − Δ S L − Δ L (35) is minimized Record Δ L 0 and Δ S L obtained from the approximation.
Uncertainties
A general discussion of the distance uncertainties is given in Clause 5
The list of uncertainties provided may not be exhaustive, and additional factors should be considered based on the specific measurement setup and procedure It is essential to utilize the mathematical framework outlined in Annex B for calculating and reporting these uncertainties.
The least-squares approximation described in section 6.4.4 utilizes the displayed distances between measurement samples to determine the distance scale deviation It is assumed that measurement samples close to L = 0 and those near the maximum location L = L max significantly impact the distance scale deviation, while samples located in the middle of the range exert less influence on the slope of the distance error model.
Applying the standard formula for the propagation of errors to Equation (4) yields the distance scale uncertainty u Δ SL in which < D otdr > ≅ D ref was used
The relationship between D otdr and D ref is approximately equal for long distances, where the standard deviation, \$u \$, represents the uncertainty in distance samples based on location samples This is analogous to the standard deviation of the slope, \$\Delta S_L\$, in the location model of Equation (10), which accounts for marker placement uncertainty and distance sampling error The least squares algorithm used to determine \$\Delta S_L\$ can also be applied to calculate \$u \$ If incremental fibers are utilized, \$\Delta L_i\$ can be averaged over the corresponding sampling interval Additionally, the uncertainty of the reference distances, \$u D_{ref}\$, can be computed using the formula \$u D_{ref}/D_{ref} = u T_b / T_b\$, where \$u T_b\$ is the uncertainty of the loop transit time, as specified in the calibration certificate of the recirculating delay line (refer to Annex A).
The location offset \$\Delta L_0\$ is determined by the intercept of the least-squares approximation on the vertical axis This intercept is primarily influenced by the initial samples near the location \$L = 0\$ and the precision of the transit time \$T_a\$.
The location offset uncertainty u Δ L0 can be calculated by applying the standard formula for the propagation of errors to Equation (34)
The uncertainty of the differences between the measured values (\$Δ L_i\$) and the least squares approximation near \$L = 0\$ is denoted as \$u_{ΔL}\$, which accounts for both marker placement uncertainty and distance sampling error This uncertainty is equivalent to the standard deviation of the differences between the measured and model values (\$Δ L_i – Δ L_{i, model}\$) near \$L = 0\$ The least squares algorithm used to determine \$Δ L_0\$ can also be applied to calculate \$u_{ΔL}\$ If incremental fibers are utilized, \$Δ L_i\$ can be averaged over the corresponding sampling interval Additionally, \$u_{Ta}\$ represents the documented uncertainty of the delay time for the lead-in fiber of the recirculating delay line, while \$u_{Ta, Θ}\$ indicates the uncertainty in delay time due to the temperature coefficient of the fiber, typically valued at 1 cm/(km °C).
The method for assessing location readout uncertainty is illustrated in Figure 2 Due to the recirculating delay line, the data generated may be insufficient to demonstrate the repetitive characteristics of the measurement samples It is advisable to identify the maximum differences between the location deviations, denoted as Δ L i (L ref), and the least-squares approximation Subsequently, this difference should be divided by the square root of 3 to calculate the location readout uncertainty, u Lreadout, which encompasses the distance sampling error.
General
The vertical scale calibration process of an OTDR for multimode fibres is divided in two parts
Loss difference calibration focuses on the receiver component of the OTDR, ensuring its capability to accurately measure the backscattered power from the fiber.
The second part involves measuring the launch conditions produced by the OTDR's laser source The goal is to characterize the near field of the OTDR source to ensure the backscatter signal is generated correctly.
The characterization of the OTDR source can be carried out using a near field analysis (see IEC 61280-1-4)
The following subclauses outline the principles of the backscattered power calibration and the near field analysis.
Loss difference calibration
Determination of the displayed power level F
For each measured loss, determine the displayed power level or an equivalent parameter that can be used to reproduce the vertical position of a measurement sample This level is termed F
Use the OTDR's clipping level at the front panel connector as the default reference point for determining F, with F ref = 0 dB All values of F should be stated in relation to this reference point; for instance, if the displayed power level is x dB below the clipping level, then F = –x dB The clipping level can be identified by introducing a sufficiently large reflection into a length of fiber, as shown in Figure 8.
Di sp la y power F ( dB)
–x dB Reference level = clipping level
Figure 8 – Determining the reference level and the displayed power level
Alternative solutions for measuring F include expressing its value in dB relative to a fixed level, provided the OTDR can display power in dB Another option is to use the starting level of the backscatter trace from a specific type of fiber at a designated pulse width as the reference level However, it's important to note that the reproducibility of this reference level can be impacted by the variability in the connection to the OTDR port.
Development of a test plan
Loss samples are influenced by power levels, distance, and the historical signal characteristics, specifically the fibre's OTDR signature before the measured feature Additionally, the detector and electronics can be impacted by the recovery from the laser's initial firing and by scattering or reflections within the fibre It's important to note that calibration is specific to the distance and signal conditions under which it is conducted.
This standard does not impose specific conditions on signal history It defines an OTDR display region A, which approximates the area where users typically take measurements Region A is characterized by four key quantities: the extrapolated start of the backscatter trace for the specific pulse width (F₀), the lowest and highest attenuation as specified in the accompanying table, and 3 dB margins on either side.
Table 2 – Attenuation coefficients defining region A
Fibre attenuation coefficients Wavelength nm Lowest ( α min ) dB/km
On the same basis, attenuation coefficient values for other wavelengths may be chosen to represent typical multimode fibres An analytical description of region A is given by dB 3 )
F max should not exceed an upper limit of 1 dB below the clipping level, unless otherwise specified by the OTDR manufacturer The loss calibration points F should lie inside region A
Calibration data for regions B and C can be submitted voluntarily Region B is relevant for fibre paths that contain components with significant loss, while Region C applies to fibre paths featuring components with substantial reflection.
Backscatter traces defined by highest and lowest fibre attenuation
Di sp la y power F ( dB)
Figure 9 – Region A, the recommended region for loss measurement samples
Develop a test plan for each method that includes sample placements, combining location and displayed power level Aim for a vertical sample spacing between 0.5 dB and 1 dB, ensuring it does not exceed the reference loss \( A_{ref} \) The attenuation range should be carefully considered.
To achieve optimal results, it is essential to select samples within region A that are evenly distributed and at a noise level of F 0 Additionally, overlapping measurement samples—those taken at the same displayed power level but from different locations—are beneficial, as illustrated in Figure 10.
D is pl ay ed power F ( dB)
Figure 10 – Possible placement of sample points within region A
Characterization of the OTDR source near field
Objectives and references
The goal is to calculate the encircle flux function EF(r) based on near-field measurements of light emitted from the end of the test cord, and to evaluate these results against the radial bound requirements.
= R r dx x xI dx x xI r EF
The encircle flux limits are defined in the IEC 61280-4-1
The requirements for the measurement and the process details are defined by the IEC 61280-
1-4 The calibration procedure is given by the IEC 61745.
Procedure
In order to be consistent with the requirements of the IEC 61280-4-1, the OTDR source near field is measured at the end connector of the test cord
Connect the end of the OTDR test cord to the near field measurement set up
Measure the encircle flux function as defined per the reference documents
Plot the measured function along with the encircled flux lower and upper bounds on a single diagram Ensure that the measured encircled flux curve is positioned between the two bounding curves.
General
Calibration methods that could provide loss calibration against known reference values are not available
The method discussed does not require prior knowledge of the loss reference, as it is assumed to be constant, leading to results that reflect only variations However, if the attenuation is known within an acceptable level of uncertainty, the measurement results can be presented in a different manner (refer to section 8.2.4).
Long fibre method
Short description
The long fibre method involves measuring the linearity of the OTDR power scale using a long multimode optical fibre Although the reference loss is assumed to be constant, its exact value remains unknown.
Equipment
The measurement equipment comprises a test OTDR and a long multimode fibre of type A1a or A1b, as specified in IEC 60793-2-10 Additionally, it includes a set of multimode lead-in fibres, which must also be of type A1a or A1b according to IEC 60793-2.
10) able to create a minimum of 1 dB attenuation; c) a variable attenuator for multimode fibre; d) mode conditioner; e) speckle scrambler optional
The attenuator and lead-in fibres are designed to establish the fibre standard at various points within region A of the OTDR display For instance, by combining three lead-in fibres with attenuation values of 2 dB, 4 dB, and 8 dB, the displayed power level can be adjusted within a range of 14 dB in 2 dB increments.
To achieve optimal performance, these fibers must feature reference type connectors, as the attenuation values account for standard connector losses For the recommended finer sample spacing of 0.5 dB, the attenuator should be adjusted in increments of 0.5 dB, ranging from 0 dB to 1.5 dB.
It should be noted that the attenuator is not necessary if the recommended steps of 0,5 dB are generated with a larger number of lead-in fibres
The mode conditioner is designed to meet launch condition requirements and prevent changes in modal conditions from affecting the fiber as the attenuator is adjusted Additionally, a speckle scrambler can be utilized to enhance repeatability.
Set of lead-in fibres
Figure 11 – Linearity measurement with a long fibre
8.2.2.2 Determination of the initial loss
The objective of this part of the procedure is to properly define reference points on the fibre to be used to always measure the same loss
To measure the total length \( D_2 \) of the long fibre, follow the OTDR manufacturer's guidelines for length measurement Select a section of the fibre, \( D_1 \), that is outside the attenuation dead zone caused by connectors in front of the fibre standard Ensure that the starting point of this section minimizes the difference \( \Delta F_{\text{max}} \) between the actual backscatter trace and its linear extrapolation, which may require the use of a lead-in fibre Finally, measure the length of the selected section \( D_1 \).
Di sp la yed power F ( dB)
Small reflection caused by connector in front of the fibre standard
Earliest start of section D 1 Start of section D 2 Δ F max
Figure 12 – Placing the beginning of section D 1 outside the attenuation dead zone
For optimal performance, a length of D 1 is suggested, which results in an initial loss of approximately 1 dB It is crucial to prevent back reflections from the far end of the fiber standard, as they can affect the accuracy of the preceding backscatter trace.
Measurement procedure
To achieve a vertical sample spacing of approximately 0.5 dB and ensure that all measurement samples are within region A of the OTDR display, it is essential to first create a comprehensive test plan for fibre and attenuator settings, as illustrated in section 7.2.2.
To minimize uncertainty in measurements, it is essential to average the loss values A_{otdr,i} across multiple displayed power levels F_{i} and over the entire length D_{1}, rather than relying on single power levels This approach is particularly beneficial at lower displayed power levels, where longer OTDR averaging can help reduce type A uncertainty It is crucial to document all applied averaging times and to record each A_{otdr,i} along with the corresponding displayed power levels F_{i} and their locations L_{i}.
Calculation and results
Calculate the non-linearity NL loss using Equation (40) max
NL ± − dB (40) where A otdr, 0 is the initial loss
Alternatively, if the attenuation of the fibre A ref is known with an appropriate level of uncertainty, calculate the loss deviation samples Δ A i using Equation (41)
Or report the loss scale deviation using Equation (42) ref ref i otdr, i
Multimode recirculating delay line for distance calibration
In this annex, a fibre-type recirculating delay line to be used as calibration artefact for multimode OTDR distance calibration is described
The device, as shown in Figure A.1, consists of a multimode four-port coupler featuring a long fiber, the length of which varies based on the calibration distance, fusion spliced between input and output ports The fiber core diameter can be either 50 μm or 62.5 μm, ensuring minimal insertion loss for long-distance measurements To maintain acceptable intermodal dispersion, the long fiber should not exceed one kilometer in length, regardless of fiber type or launching conditions Additionally, an input fiber, typically around 250 m, is spliced to the second input port and equipped with a connector, while a short output fiber, less than 1 m, is spliced to the second output port and terminated with a low back-reflection connector Finally, a high reflectance device is utilized to create a reflection at the end of the output fiber.
The lead-in fibre length is determined by the input fibre, one branch of the coupler, and the output fibre Meanwhile, the loop length is defined by the long fibre mentioned in item a) and the second half of the coupler.
Length of lead-in fibre
The goal is to calibrate a recirculating delay line by focusing on two key parameters: the loop's fibre transit time, denoted as \$T_b\$, and the input length transit time, represented as \$T_a\$ The input length transit time \$T_a\$ is the total of the transit times for the input fibre, coupler pigtails, and output fibre.
The setup includes a pulse generator with adjustable rate and delay, a digital counter, electro-optical (E/O) and optical-electrical (O/E) converters, an oscilloscope, and an optical attenuator It is essential to know the center wavelength of the E/O converter.
To determine the loop transit time, set up the system as shown in Figure A.2, utilizing a mode conditioner (MC) to create a launch condition similar to that produced by the OTDR or to generate an overfill Adjust the pulse generator for suitable pulse width and the attenuator for appropriate amplitude for the E/O converter Set the pulse repetition rate to 200/L b kHz, where L b is the loop's fibre length in kilometres Observe the output pulses from the O/E converter on the oscilloscope and adjust the repetition rate until the oscilloscope trace displays two superimposed pulses: the optical pulse transmitted directly from the E/O converter and the fraction of the pulse that has traveled once around the loop Record the repetition period with a digital counter, which represents the loop transit time T b Finally, estimate the uncertainty u T b by making small adjustments to the repetition rate.
Figure A.2 – Measurement set-up for loop transit time T b A.2.3.2 Lead-in transit time measurement
To calibrate the transit time of lead-in fibres, set up the system as shown in Figure A.3, utilizing a mode conditioner (MC) to create a launch condition similar to that of the OTDR or to generate an overfill Adjust the pulse generator's pulse width, amplitude, and repetition rate to approximately 1 kHz, and provide a trigger signal for the digital counter It may be necessary to delay the output pulse to the E/O converter relative to this trigger pulse due to the "dead time" in some counters Set the trigger level of channel B on the counter to ensure it is reliably triggered by the O/E converter's output, avoiding false triggers from smaller pulses Record the transit time T1 displayed by the counter, then temporarily disconnect the O/E converter from the counter and connect it to an oscilloscope to view and note the amplitude of the directly transmitted pulses.
Remove the recirculating delay line and connect the optical-to-electrical (O/E) converter directly to the attenuator Observe the pulses on the oscilloscope and adjust the attenuator to match the previous pulse amplitude Reconnect the O/E converter to the counter to measure the transit time \( T_2 \) The transit time of the lead-in fiber \( T_a \) can then be determined.
Figure A.3 – Calibration set-up for lead-in transit time T a A.3 Uncertainties
The guidelines of the mathematical basis given in Annex B should be used to calculate and state the uncertainties
The total uncertainty of the loop transit time, denoted as \( u_{T_b} \), is determined by accumulating various contributions through root-sum-squaring These contributions include \( u_{T_b,\text{counter}} \), which represents the time uncertainty from the digital counter due to clock frequency uncertainty and time interval resolution; \( u_{T_b,\text{adjust}} \), the time uncertainty arising from adjusting the repetition rate using an oscilloscope; \( u_{T_b,\lambda} \), the time uncertainty linked to the E/O converter's center wavelength, calculated by multiplying the wavelength uncertainty by the fiber length \( L_a \) and the chromatic dispersion; and \( u_{T_b,\Theta} \), the time uncertainty associated with the fiber's temperature coefficient, which has a typical value.
1 cm/(km °C) within the allowable temperature range; u Tb,MD the time uncertainty due to the modal dispersion at the applied launch condition
The total uncertainty of the lead-in transit time, denoted as \( u_{T_a} \), is determined by accumulating various contributions through root-sum-squaring This includes \( u_{T_a,\text{counter}} \), which represents the time uncertainty from the digital counter due to clock frequency uncertainty, time interval resolution, and trigger amplitude settings Additionally, \( u_{T_a,\text{type A}} \) accounts for type A time uncertainty, such as timing jitter, which can be derived from successive counter readings Furthermore, \( u_{T_a,\lambda} \) reflects the time uncertainty related to the E/O converter's center wavelength, calculated by multiplying the wavelength uncertainty by the fiber length \( L_a \) and the chromatic dispersion Lastly, \( u_{T_a,\Theta} \) represents the time uncertainty due to the fiber's temperature coefficient, with typical values provided.
1 cm/(km °C) within the allowable temperature range; u Ta,MD the time uncertainty due to the modal dispersion at the applied launch condition
Additional contributions may have to be taken into account, depending on the measurement set-up and procedure
The calibration results for the recirculating delay line must include the approximate lengths of the lead-in fiber and loop, the measured transit times for both, the centroidal wavelength of the electro-optical (E/O) converter, and the time uncertainties of ±2 u Ta and ±2 u Tb, as specified in Clause A.3.
This annex summaries the form of evaluating, combining and reporting the uncertainty of measurement It is based on the "Guide to the expression of uncertainty in measurement”
(ISO/IEC Guide 98-3(GUM) It does not relieve the need to consult this guide for more advice
This standard identifies two methods for evaluating measurement uncertainty Type A involves statistical analysis of multiple measurements of the same measurand, while Type B relies on other relevant knowledge to assess uncertainty.
The type A evaluation of standard uncertainty can be applied when several independent observations have been made for a quantity under the same conditions of measurement
For a quantity X estimated from n independent repeated observations X k , the arithmetic mean is
This mean is used as the estimate of the quantity, that is x = X The experimental standard deviation of the observations is given by
X is the arithmetic mean of the observed values;
X k are the measurement samples of a series of measurements; n is the number of measurements; it is assumed to be large, for example, n ≥ 10
The type A standard uncertainty u typeA (x) associated with the estimate x is the experimental standard deviation of the mean n
Type B evaluation of standard uncertainty assesses uncertainty through methods other than statistical analysis of observations This approach relies on scientific judgment and considers all available information regarding the variability of the quantity in question.
When estimating a quantity \( X \) from a manufacturer's specification or calibration certificate, if the quoted uncertainty \( U(x) \) is expressed as a multiple \( k \) of the standard deviation, the standard uncertainty \( u(x) \) can be calculated by dividing the quoted value by the multiplier This relationship is represented by the formula \( u(x) = \frac{U(x)}{k} \).
When only the upper limit (X max) and lower limit (X min) can be estimated for a quantity X, such as in a manufacturer's specifications or a temperature range, a rectangular probability distribution is assumed The estimated value of X is derived from this distribution.
1 X max X min x= + (B.5) and the standard uncertainty is
The standard uncertainty in the output estimate \( y \) is influenced by the standard uncertainty in the input estimate \( x \), expressed as \( u(y) = c x u(x) \), where \( c \) represents the sensitivity coefficient This coefficient is defined as the partial derivative of the model function \( y(x) \), evaluated at the input estimate \( x \).